CN112651465B - Equipment target interpretation method based on high-resolution remote sensing image - Google Patents

Abstract

The invention discloses a device target interpretation method based on a high-resolution remote sensing image, which is characterized in that full-color spectrum images and multispectral images in an original remote sensing image are fused, preprocessed, and further adjusted and screened to obtain a target analysis high-resolution remote sensing image; establishing an equipment target classification system according to the equipment target types; directly interpreting according to the recognition characteristic information of the reaction on the target analysis high-resolution remote sensing image to realize the preliminary interpretation and recognition of the target; indirectly interpreting the target which cannot be determined through a logical reasoning method, and realizing the secondary confirmation interpretation and identification of the target to obtain an interpretation result; and verifying the interpretation result of the equipment target by a contradiction analysis method, and analyzing and identifying the target. A systematic high-resolution remote sensing image equipment target interpretation method is established, and interpretation of equipment targets in the existing interpretation theory is perfected.

Description

Equipment target interpretation method based on high-resolution remote sensing image

Technical Field

The invention relates to the technical field of remote sensing image interpretation, in particular to an equipment target interpretation method based on a high-resolution remote sensing image.

Background

The equipment targets comprise military aircraft, ships and land vehicle equipment, are the basis of army construction, are important weapon strength performance in military operations, and are important striking objects of modern war including military reconnaissance, monitoring and guidance. Remote sensing technology has become an indispensable technology in modern war today, and is an effective means for searching, detecting and predicting changes of military targets and executing the changes. With the continuous development of remote sensing technology, the application level is difficult to look at, and the root is that the knowledge system is imperfect for the interpretation method of the identification target, and particularly, a plurality of problems still exist in the interpretation method of the equipment target of the remote sensing image: (1) Few and no systematic guidance for equipment target interpretation methods; (2) The method has the problems of interpretation method and actual application, and weak actual application; (3) The existing theoretical method is asynchronous with the development of the remote sensing image technology, and the problems of rapid updating development of the remote sensing technology and slow development of the theoretical guiding method exist.

Therefore, how to implement the equipment target interpretation adapted to the rapidly developed remote sensing technology is a problem that needs to be solved by those skilled in the art.

Disclosure of Invention

In view of the above, the present invention provides a method for interpreting equipment targets based on high-resolution remote sensing images, which uses a logical reasoning method to determine the steps of the method for interpreting equipment targets based on high-resolution remote sensing images, so as to realize the operability of interpreting equipment targets and realize the synchronous development of interpretation methods in cooperation with the updating of remote sensing image technology.

In order to achieve the above purpose, the present invention adopts the following technical scheme:

an equipment target interpretation method based on high-resolution remote sensing images comprises the following steps:

step 1: the method comprises the steps of obtaining an original remote sensing image, preprocessing a full-color spectrum image and a multispectral image in the original remote sensing image, and then fusing to obtain a high-resolution remote sensing image;

performing radiation calibration, atmospheric correction and orthographic correction pretreatment on the multispectral image, and performing the radiation calibration and orthographic correction pretreatment on the full-color spectral image;

step 2: screening the high-resolution remote sensing images according to a set rule, and adjusting the remote sensing images which do not meet the rule to obtain a target cloud-free covered and clear target analysis high-resolution remote sensing image;

in order to obtain a high-resolution remote sensing image capable of performing visual interpretation of a target, the fused high-resolution remote sensing image is screened according to a set rule, wherein the rule comprises:

(1) The spatial resolution of the image of the research area is better than 0.5 meter;

(2) The target of the research area is free of cloud coverage, namely the cloud coverage amount is less than or equal to 1%;

(3) The exposure rate and the chromatic aberration of the image meet the requirement of visual interpretation;

(4) The side swing angle of the satellite is not more than +/-15 degrees during image acquisition;

if the rule is not satisfied, the fused high-resolution remote sensing image needs to be correspondingly adjusted, wherein the adjustment comprises image mosaic, image clipping, image band combination and adjustment of exposure rate and color by using image processing software such as ENVI, PS and the like;

step 3: establishing an equipment target classification system according to the types of the equipment targets; the equipment target classification system comprises target type classification, target model classification and target feature information, wherein the target feature information corresponds to the target type and the target model; the target type classification comprises a classification system of military aircraft, ships and ground equipment;

step 4: extracting characteristic information of the equipment target on the target analysis high-resolution remote sensing image, wherein the characteristic information comprises characteristics of the equipment target, common sense information related to the equipment target and the like, selecting identification characteristics from the characteristic information, comparing the identification characteristics with the equipment target classification system, and directly interpreting the equipment target to obtain a target primary identification result;

aiming at characteristic information reflected by the equipment target on the target analysis high-resolution remote sensing image, six identification features are selected from the characteristic information, and the equipment target is directly interpreted based on the identification features, so that preliminary interpretation and identification of the target are realized; if the identification features are clear and obvious, the direct interpretation can directly obtain the equipment target model information, or only can obtain the equipment target large-class information, the equipment target type and the like because the identification features are less;

the characteristic information comprises the characteristics and conditions of the target, which are reflected by the target image of the equipment target in terms of the shape, size, tone, shadow, position, activity and the like, and is used as the basic basis for identifying and judging the characteristics and conditions of the target, and the characteristics are interpreted as the identification characteristics of the equipment target;

step 5: extracting image information on the target analysis high-resolution remote sensing image, wherein the image information comprises image information of the equipment target and image information except the equipment target, secondary interpretation identification features are selected from the image information, surrounding related event information is obtained, the surrounding related event information is related event information of the image information, and according to a primary target identification result, the secondary interpretation identification features, the equipment target classification system and the surrounding related event information, the secondary interpretation identification of the equipment target is realized through a logical reasoning method, and a target model is obtained;

step 51: extracting the image information on the target analysis high-resolution remote sensing image, and carrying out primary classification according to the target primary identification result to obtain a target major type, namely determining a target equipment major type, belonging to an airplane, a ship or ground equipment;

the first class classification of the object to be interpreted can be determined according to the six identification features; the method comprises the steps that an airplane is parked at an airport, a ship is parked at a port, vehicles of ground equipment are parked at land, and the targets are classified in one stage based on different states of three targets of the airplane, the ship and the ground equipment on the image based on a high-resolution image;

step 52: extracting the secondary interpretation recognition features from the image information, carrying out secondary classification by combining the secondary interpretation recognition features, the target major class type and the equipment target classification system, and reasoning to obtain a secondary attribute classification result;

the secondary interpretation recognition features comprise position information of the target analysis high-resolution remote sensing image, a secondary classification result is deduced according to the position of the target, and whether the target is military or civil is judged;

the method for determining the types of military and civil is divided into two types, wherein the first type is that information such as numbers in a target analysis high-resolution remote sensing image can be extracted directly according to the characteristics of the target, if the military numbers exist, the military is the military, or the target types are determined in a target primary identification result, and the military type or the civil type can be determined according to the types of aircrafts, ships and ground equipment; secondly, if the target cannot be directly confirmed, judging by means of the position characteristics of the equipment target, acquiring the position characteristics of the image of the target, wherein the position can be positioned through longitude and latitude, and the target parked at the military base is of a military type under the general condition;

step 53: acquiring the surrounding related event information related to the image information, and reasoning to acquire the target model according to the target major type, the secondary attribute classification result, the equipment target classification system and the surrounding related event information;

the surrounding related event information includes, for example, in which country the current position is located, which equipment library, the type, model, etc. of the existing equipment in the equipment library; the equipment target classification system stores definition characteristics corresponding to target types, such as various components of various aircraft targets, materials and materials of the components, a power system of an aircraft and the like; comparing and reasoning the large class type, attribute, surrounding related information and the like with storage characteristics in the equipment target classification system to determine a specific model;

step 6: and (3) acquiring parameter indexes and target related event information of the equipment targets corresponding to the target model, comparing the parameter indexes and the target related event information with the characteristic information and the image information of the equipment targets, searching contradiction points, returning to the step (5) if the contradiction points exist, and otherwise, outputting the current target model.

Compared with the prior art, the invention discloses an equipment target interpretation method based on a high-resolution remote sensing image, which is used for preprocessing full-color spectrum images and multispectral images in an original remote sensing image; then fusing the preprocessed full-color spectrum image and the preprocessed multispectral image, and further adjusting and screening to obtain a target analysis high-resolution remote sensing image; establishing an equipment target classification system according to the equipment target types; directly interpreting according to the recognition characteristic information of the reaction on the target analysis high-resolution remote sensing image to realize the preliminary interpretation and recognition of the target; indirectly interpreting the undetermined target by combining various characteristic information and extracted influence information around the target through a logical reasoning method to realize secondary confirmation, interpretation and identification of the target and obtain the model of the target; and verifying the interpretation result of the equipment target by a contradiction analysis method, and analyzing and identifying the target. The method for interpreting the equipment targets by using logical reasoning and repeated verification to form systematic high-resolution images achieves unified system classification of the equipment targets including military aircraft, ships and ground equipment, achieves target interpretation of the equipment targets by using the logical reasoning method and the contradiction analysis method, and perfects interpretation of the equipment targets in the existing interpretation theory.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flowchart of an equipment target interpretation method based on a high-resolution remote sensing image provided by the invention;

FIG. 2 is a diagram showing the classification of an aircraft target system according to the present invention;

FIG. 3 is a diagram showing the classification of a ship target system according to the present invention;

FIG. 4 is a diagram showing a classification of a ground equipment target system provided by the invention;

FIG. 5 is a schematic diagram of a remote sensing image according to embodiment 1 of the present invention;

FIG. 6 is a schematic diagram of a remote sensing image according to embodiment 1 of the present invention;

fig. 7 is a schematic image diagram of embodiment 2 provided in the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

The embodiment of the invention discloses a device target interpretation method based on a high-resolution remote sensing image, which comprises the following steps of:

s1: the method comprises the steps of obtaining an original remote sensing image, preprocessing a full-color spectrum image and a multispectral image in the original remote sensing image, and then fusing to obtain a high-resolution remote sensing image; fusing the preprocessed full-color spectrum image and the preprocessed multispectral image by adopting ENVI software;

performing radiation calibration, atmospheric correction and orthographic correction pretreatment on the multispectral image, and performing radiation calibration and orthographic correction pretreatment on the full-color spectral image;

s2: in order to obtain a target analysis high-resolution remote sensing image capable of performing target visual interpretation, the fused high-resolution remote sensing image is screened according to a set rule, wherein the rule comprises:

(1) The spatial resolution of the image of the research area is better than 0.5 meter;

(2) The target of the research area is free of cloud coverage, namely the cloud coverage amount is less than or equal to 1%;

(3) The exposure rate and the chromatic aberration of the image meet the requirement of visual interpretation;

(4) The side swing angle of the satellite is not more than +/-15 degrees during image acquisition;

if the rule is not satisfied, the fused high-resolution remote sensing image needs to be correspondingly adjusted, wherein the adjustment comprises image mosaic, image clipping, image wave band combination and exposure rate and color adjustment by utilizing image processing software Photoshop software such as ENVI, PS and the like;

s3: establishing an equipment target classification system according to the types of the equipment targets, wherein the equipment target classification system comprises target type classification, target model classification and target feature information, and the target feature information corresponds to the target types and the target models; the target type classification comprises a classification system of military aircraft, ships and ground equipment;

according to the interpretation objects, including planes, ships and ground equipment, according to the characteristics of functions, performance actions and the like, classifying the targets of the planes, the ships and the ground equipment, and classifying the military planes in the planes, the military ships in the ships, the military vehicles in the ground equipment targets and the vehicle-mounted missiles in detail, so as to establish a corresponding target classification system;

s4: extracting characteristic information of equipment targets on the high-target analysis resolution remote sensing image, wherein the characteristic information comprises characteristics of the equipment targets, common sense information related to the equipment targets and the like, selecting identification characteristics from the characteristic information, comparing the identification characteristics with an equipment target classification system, and directly interpreting the equipment targets to obtain a target primary identification result;

aiming at the characteristic information of the equipment target reflected on the target analysis high-resolution remote sensing image, six identification features are selected from the characteristic information, and the equipment target is directly interpreted based on the identification features, so that the preliminary interpretation and identification of the target are realized;

the characteristic information comprises the characteristic information of the equipment target, wherein the characteristic information comprises the characteristic and the condition of the equipment target reflected by the target image according to the shape, the size, the tone, the shadow, the position, the activity and other phenomena, and is used as the identification characteristic of the equipment target on the basis of identifying and judging the characteristic and the condition of the target, and the identification characteristic is expressed as follows:

the shape refers to the reflection of the external outline of the ground object and the shape of the detail condition on the image; different types of ground targets have specific shapes, are important basis for identifying the targets, are different from images seen by eyes, and are characterized in that the eyes are always in head-up view, the side surfaces of the objects are seen, the remote sensing satellite images are photographed from above the objects, the top shapes of the main objects are reflected, but the shapes of the targets reflected in the air are also influenced by many factors such as relief, object height, photographing angle and the like, and the situation that shape features are camouflaged, such as camouflage nets are covered on the vehicle targets, can also occur;

size refers to the dimension of the ground object on the image film, such as length, width, area, volume, etc.; the size characteristics of the ground object mainly depend on an image scale, the size connection of the object and the image can be established according to the image scale, the actual size of the object is calculated and obtained, and the size of the object is influenced by topography fluctuation and shooting inclination angle;

tone, which is the reflection of electromagnetic wave characteristics of an object on an image, is presented as color on a fused high-resolution remote sensing image; the shape and the size of the ground object are displayed through color tones, so that the color tone characteristic is also the most basic interpretation mark of the color characteristic;

shadow refers to a shadow generated by an object which is higher than or concave on the ground under the irradiation of direct sunlight and is divided into a principal shadow and a falling shadow;

the position refers to the reflection of the position where the ground object exists and the spatial position relation among the ground objects in the image, and reflects the relation among targets, which is an important indirect interpretation feature;

activity refers to various symptoms caused by the activity of a target, and any target can generate symptoms of the activity as long as the target has the activity, and the symptoms have a certain relation with the property of the target.

Based on the image interpretation recognition of six recognition features, the direct judgment of the target is carried out according to the detailed image recognition features of the airplane, the ship and the ground equipment, and the recognition features of the airplane, the ship and the ground equipment are as follows:

the aircraft is identified on the image, so that the aircraft is generally easy to interpret, and the type, the number and the state of the aircraft are specified, and the condition of the aircraft carrying the nuclear weapon is interpreted; different types of aircrafts have the characteristics that the types of aircrafts are interpreted on images according to the shapes, the number, the installation positions and other positions of all the components of the aircrafts, and the influences on the overall shape of an aircraft target are different based on the identification characteristics, particularly the shapes and the material characteristics, of the basic components of the aircrafts, such as wings, power devices, tail wings, fuselages, landing devices, weapons and the like, and the shape characteristics of all the components reflected on the images are different;

the ships and the interpretation of the ships generally need to interpret the types of the ships, distinguish the ship levels and different types of the ships, and have different characteristics in tactical and technical performance aspects; on the image, various ships are identified according to the characteristics of the various ships;

ground equipment, which is used for mainly interpreting military vehicles; military vehicles are the collective names of various automobiles, tracked vehicles and motorcycles of army equipment, and are generally reflected on images to be in rectangular blocks; the vehicles have different application functions, different vehicle forms and different images, for example, the command vehicles are usually in smaller rectangular blocks on the images, the differences between the color tone of the carriage of the jeep with the roof and the color tone of the headstock are obvious, the jeep without the roof is arranged, and the images of the carriage part are disordered; the color tone of the minibus and the sedan is basically consistent; the command vehicles on the images are easy to distinguish, such as tracked vehicles, and the tracked vehicles are mainly divided into armored conveying vehicles and infantry combat vehicles, so that the armored conveying vehicles have good maneuvering performance and strong protection capability, the top of the armored conveying vehicles is flat and clear, the edges and corners of the surfaces are obvious, the images are rectangular blocks, the track marks are obvious, and the tracked vehicles are similar to tanks but have no gun towers; the gun turret is a main basis for distinguishing the armored transport vehicle from the self-propelled gun and the tank, in addition, the armored transport vehicle is mostly located at a hidden position on the position information, and the tank and the self-propelled gun are mostly occupied at a terrain which is favorable for fire development; the infantry war chariot is mainly used for cooperative tank combat, can also independently do tasks and is generally amphibious; on the image, the infantry war chariot with only the machine gun and the machine gun is very similar to the armored conveying vehicle, and compared with the self-propelled gun, the infantry war chariot with the gun is slightly smaller in gun tower and shorter in gun barrel; in addition, for more common military engineering vehicles, medical ambulances, liquid tank trucks, automobiles, motorcycles and the like, the special purpose is achieved, and certain special points are necessarily provided on the appearance structure, for example, a working room of the military engineering vehicle is mostly a firm metal box car, wherein the automobiles are provided with hoisting equipment such as cranes and the like, and the equipment with special functions is the basis for image interpretation;

s5: extracting image information on a target analysis high-resolution remote sensing image, wherein the image information comprises image information of an equipment target and image information except the equipment target, selecting secondary interpretation recognition features from the image information, acquiring surrounding related event information which is related event information of the image information, and indirectly interpreting the image information according to a target primary recognition result, the secondary interpretation recognition features, an equipment target classification system and the surrounding related event information by a logical reasoning method to realize secondary interpretation recognition of the equipment target, so as to obtain a target model;

the logical reasoning method is an interpretation method for combining objective evidence and real state of the object according to the general rule of the object development, forming concepts according to the rule and rule of certain logical thinking, and judging and reasoning by using a determination method and an elimination method; aiming at S4, directly decoding a target model of the equipment target, and indirectly decoding the target model through a logical reasoning method, the specific process is as follows:

s51: extracting image information on the target analysis high-resolution remote sensing image, and carrying out primary classification according to a target primary identification result to obtain a target large class type;

performing primary classification on the target image on the target analysis high-resolution remote sensing image according to the identification characteristics:

firstly, determining first-class classification of an object to be interpreted according to six identification features; the method comprises the steps that an airplane is parked at an airport, a ship is parked at a port, vehicles of ground equipment are parked at land, and the targets are classified in one stage based on different states of three targets of the airplane, the ship and the ground equipment on the image based on a high-resolution image;

s52: extracting secondary interpretation recognition features from the image information, carrying out secondary classification by combining the secondary interpretation recognition features, the target major class type and the equipment target classification system, and reasoning to obtain a secondary attribute classification result;

combining the position characteristics of the equipment targets to carry out secondary classification:

the military and civil type is determined. The military or civil model cannot be judged directly generally, the position characteristics of the equipment target are needed to be judged, the target parked on the military base is a military type in general, and the classification and the position of the target can be further determined by determining the military or civil model; the position may be located by latitude and longitude.

S53: deducing a grading major class of the target according to S51-S52, acquiring surrounding related event information related to the image information, and deducing and acquiring a target model according to the type of the major class of the target, a secondary attribute classification result, a device target classification system and surrounding related event information;

grasping the type of the large class, such as determining that the first class classification result targets are planes; mastering the definition of the airplane major class according to the equipment target classification system, and mastering each component part of the airplane target, the materials and the materials of each component part, the power system of the airplane and the like; and determining a specific target model by comparing the defined characteristics in the equipment target classification system according to the target major type, the attribute and the surrounding related event information.

Aircraft target:

firstly, acquiring the actual size of an airplane, including physical parameter information such as the size of the airplane body, the proportional position of wings and the airplane body and the like; acquiring a composition structure of a machine body; the method comprises the steps of obtaining the forms of all the constituent structures, such as an upper single wing, a middle single wing and a lower single wing, wherein the shapes of the wing are triangular, trapezoidal, semicircular and the like; acquiring the type, the position and the shape of an engine of an aircraft; acquiring coating color of an airplane body and text and graphic information printed on the airplane body; matching the acquired parameter information with a target image in the image to exclude targets which do not accord with the parameter;

according to the position information, targets such as airplanes in the United states cannot appear in the Japanese air force base, all the airplanes of each organization have the air force base and the position, and other options are excluded according to the information;

ship target:

firstly, acquiring the actual size of a ship, including physical parameter information such as ship length, width, length-width ratio and the like; acquiring a composition structure of a ship; the method comprises the steps of obtaining the forms of each component structure, such as the positions of ships and warships, the number of chimneys, the number and shape of smoke outlets, the existence of helicopter decks and the like; acquiring digital, text and graphic information printed on special facilities of ships, such as ships, cranes and ship bodies;

military ships are generally parked in fixed ports, cross parking of the ships between countries is less, the ships not belonging to the ports are eliminated according to the position information, and attribution information is determined;

ground equipment:

firstly, acquiring the actual size of a vehicle, including the physical parameter information of the length, width, length-width ratio and the like of the vehicle; acquiring a composition structure of a vehicle, wherein the vehicle is provided with or not with auxiliary facilities such as a crane and the like; the method comprises the steps of obtaining information such as painting of a vehicle and special symbols of the vehicle, for example, symbols of a red cross of a medical ambulance;

the ground equipment has certain position characteristics, and the coating colors and shapes of the ground equipment in each country and region are different;

s6: acquiring parameter indexes and target related event information of equipment targets corresponding to the target model, comparing the parameter indexes and the target related event information with characteristic information and image information of the equipment targets to find contradiction points, returning to S5 if the contradiction points exist, and otherwise outputting the current target model;

verifying and verifying the interpretation result of the equipment target by a contradiction analysis method, and analyzing and identifying the target;

the contradiction analysis method is to grasp things from the inherent relation of things, analyze things from the contradictions of things and solve the objectively existing problems; the remote sensing interpretation of the equipment target is to continuously reveal the nature of things from the contact and contradiction of the ground object reflected by remote sensing, and the actual information of the target is verified through the contact and contradiction relation among the things;

for high-resolution remote sensing image interpretation, points which do not accord with actual conditions are contradictory points, any contradictory point is necessarily provided with a break, and the break is always the key point for finding the true interpretation, for example, for a ground equipment vehicle with branch camouflage, the interpretation difficulty is increased because a camouflage pattern is very similar to a target background, but camouflage such as branches is usually limited by seasons, areas and equipment, the best is difficult to achieve, and after a period of time, camouflage effects such as leaf withering and showing a target appearance are reduced by some camouflage materials, so that the targets in different time phases are observed according to the image difference in time phases, the difference between the targets and surrounding matters is noted, the decoy is eliminated, and the nature of the targets is determined by utilizing the contradictory point.

In order to further optimize the above technical solution, if the target analysis high-resolution remote sensing image in S5 includes only the equipment target image, then the target image is subjected to depth analysis interpretation, according to the initial recognition result of the target of the equipment target, relevant events of the equipment target are collected, relevant information including hundreds of degrees encyclopedia, news, announcements issued by the official network, etc. is collected through the network, and the target is subjected to depth analysis reasoning interpretation in combination with the relevant events of the target.

In order to further optimize the technical scheme, in S5, the preliminary identification result of the target in S4 is simultaneously verified according to more extracted image information and related event information.

Example 1

As shown in fig. 5-6, in view of the high-resolution remote sensing image, basic parameters of the target are determined, that is, feature information of the equipment target is obtained, and the information is shown in the following table 1:

TABLE 1 direct interpretation of the obtained target basic parameter information

According to the above features, it can be first determined that the installation target in fig. 5 is an aircraft, model number C-130;

secondly, according to the information interpreted on the remote sensing image map in the table 1, the information is indirectly interpreted by using logical reasoning, and the following information is determined:

(1) Obtaining information according to the national flag value with red round symbols of Japan on the target wing reflected by the image: the country of interest of the aircraft is japan;

(2) Obtaining information of the large aircraft according to parameters such as the length and the like obtained by direct interpretation; according to the target classification judgment shown in fig. 2-4, combining the information of the type of aircraft being a large aircraft to obtain information: typically a transport, a fuel dispenser, or a service aircraft;

(3) According to the form of the wing span of the target wing reflected by the image, information is obtained: the aircraft is a fixed wing aircraft;

(4) From the direct interpreted content and the indirect interpreted content: the aircraft shown in fig. 5 and 6 has the same wing shape, tail wing and other components, the same engine, auxiliary fuel tank and other components, and the same number of positions, and information is obtained: the functions and the purposes of the two are similar;

more image information and related event information are acquired for further indirect interpretation, if only equipment targets are arranged on the image, depth analysis is carried out on the equipment targets, and information reflected by the target images is obtained: the only difference is that fig. 6 has a fueling module, which gets information based on common knowledge of the aircraft: the purpose of the oiling cabin is oiling, and the aircraft in FIG. 6 is proved to be an oiling machine;

meanwhile, the direct interpretation result is verified:

according to the image response content, the basic information such as the airframe, the airframe length, the number of engines and the like, the model number of the airplane is determined to be C-130, and then the model number is indirectly interpreted to be verified:

parsing deep feature information of the target is: the wing is in the form of a front straight upper single wing, trapezoid horizontal displacement is positioned at the root of a vertical tail wing, an engine position and an airport where the trapezoid horizontal displacement is positioned are Japanese airports, the information in the fifth step is determined to be completely corresponding to C-130 by comparing live-action photos with the gathering correspondence of Japanese aircraft equipment data, and the model machine is verified to be C-130;

(5) According to the understanding of the aircraft equipped in japan, two similar aircraft with different uses are obtained in the aircraft equipment in japan, the aerial photographs of the two aircraft and the relevant guidance of the aircraft equipment in japan are searched, the KC-130 is provided with an externally hung refueling bin, the targets in fig. 5 are not removed, the targets in fig. 5 are KC-130, and the targets in fig. 5 and fig. 6 are respectively the C-130 conveyor and the KC-130 refueling machine by verifying that the relevant data of the C-130 are matched with the characteristic information extracted from the images.

(6) Then, the purpose of the two types of machines is determined by a contradictory analysis method:

although the appearance of the two types of machines on the remote sensing image is very similar, the functional properties of the two types of machines are basically changed due to the existence of the oil adding cabin, the figure 5 without the oil adding cabin is a conveyor, the figure 6 with the oil adding cabin is an oiling machine, further analysis is carried out, the two types of machines are different in the fight task due to the change of the properties, the fight sequence during fight is also different for the fight target, the oiling machine is the primary fight target, and the conveyor is the secondary.

Finally, the transporter C-130 of FIG. 5 is interpreted from the image of the high resolution remote sensing satellite, and is not the primary target for striking during striking, and the fuel dispenser KC-130 of FIG. 6 is the primary target for striking during striking.

The method is characterized in that the method comprises the steps of interpreting the high-resolution remote sensing satellite images, and finally, the target is interpreted uniformly in terms of judging the type, the function and the battle of the airplane, so that a result is obtained.

Example 2

FIG. 7 is a photograph of the ground of a team in the 4 th early warning machine of the Bastein air force when the team arms at the Ma Suer air force base.

1. Logical reasoning process:

1. first, physical information is observed.

The image information of the direct reaction in the photo has at least 3 aspects:

(1) Location: buckstan Ma Suer air force base;

(2) Early warning machine: is one of 4 ZDK-03 early warning machines introduced from China.

(3) The hangar: the arch system structure is adopted, 7 arch beams are arranged, 7 rear wall upright posts are arranged, and the aircraft hangar is a large building special for aircraft maintenance and parking.

2. From the above information, 4 decisions can be made:

(1) Type of hangar: the ZDK-03 early warning machine library; judging the basis: the ZDK-03 early warning machine is parked in the hangar.

(2) Hangar size: the ZDK-03 type early warning machine library can be judged to be at least longer than 38 meters and wider than 42 meters; judging the basis: the ZDK-03 early warning machine takes a nine-transporting machine as a bearing platform, the length of the nine-transporting machine is 34 meters, the span of the nine-transporting machine is 38 meters, and the parking safety distance of a large-sized airplane in a hangar is required to be more than 2 meters.

(3) Shape of the hangar: the machine base is square, and the top of the machine base is a slope herringbone; judging the basis: the top of the arched girder has a certain included angle, the lower part of the arched girder is a hangar side wall support, and the upright post is a hangar rear wall support.

(4) Hangar location: pakistan Ma Suer air force base.

The image analysis of the Ma Suer air force base shows that the base is internally provided with a base 3 seat which is longer than 38 meters and wider than 42 meters and has a herringbone slope surface at the top. The hangar parking apron is parked with a ZDK-03 early warning machine 1 frame. The preliminary target recognition result of direct interpretation is as follows: and (5) parking the ZDK-03 early warning machine library for the Baair force.

The process of direct interpretation: the ground photo information is converted into 4 recognition features of the hangar, so that the connection among the early warning machine, hangar building materials and the hangar is tightly grasped, and the hangar type is deduced according to the position of the early warning machine; deducing the position of the hangar by using the photo shooting place; deducing the size of the hangar according to the size of the early warning machine; the shape and the whole shape of the top of the hangar are deduced by the arched beams and the upright columns of the hangar.

Indirect interpretation: there is always contradiction between things, and indirect reasoning and verification are carried out on the preliminary conclusion directly interpreted;

searching contradiction points: preliminarily judging that the ZDK-03 pre-warning machine base of the Bakistan air force is influenced by the characteristics of remote sensing images, whether the internal structure accords with the characteristics reflected by the photos is not confirmed, and the number of 3 machine bases is different from that of 4 known ZDK-03 pre-warning machines introduced by the Bakistan, so that contradiction occurs between the number of the machines and the number of the machine bases;

acquiring target related event information, and knowing: in 2012, in one attack of the Baskistan-gambler air army base, 1 Sabo-2000 Aili early warning machine is scrapped, and the ZDK-03 early warning machine is not damaged. This message shows that the ZDK-03 pre-warning machine is deployed at the gahler air force base.

And then turning over satellites under different images of the site to obtain remote sensing images of the Pickle air force base of 2016, 10 months and 3 days, wherein 1 base of the base is built with the same-size and same-shape hangars of Ma Suer air force base of 3 bases, and meanwhile, the remote sensing images of 2014, 10 months and 4 days obtain the situation that the hangars are built, so that 7 arched beams are found in total in the hangars, and 7 upright posts are found in the rear wall. The structure is identical to the photo display information.

Through the remote sensing image analysis, the hangar is a special hangar for parking the early warning machine. However, in the image interpretation process, the base is found to have a Sabo-2000 'Aili' early warning machine stopped, which is consistent with the report of attack of the previous bus early warning machine, but the stop early warning machine is not matched with the hangar.

Determining contradiction points:

in total, 2 contradictions occur: ma Suer the number of the air force base early warning machines is inconsistent with the number of the special hangars; the types of the wall air force base early warning machine base and the resident early warning machine are inconsistent.

The model ZDK-03 of the early warning machine base of the Baskistan is primarily determined through the relation and contradictory relation among things, but the problem that the number and the types of the early warning machines are inconsistent with the machine base still exists. All air force bases in the pakistan are analyzed, and the 6 air force bases of Mi Yangwa, raffmyl, lissal, quadax, nanowatt Bo Shaa, mu Shafu pakistan and the like are found to be respectively built with libraries with the same size and shape. It can be determined that the baair force has such a hangar 10.

Searching the connection between the 10-seat hangar and the Pakistan warning machine:

the network information is comprehensively known: ' salbo-2000 ' ai Liye ' introduced from Sweden in 2006 is provided with 4 frames of an air early warning machine; 4 ZDK-03 early warning machines are introduced from China in 2010; the 1 st Sabo-2000 'Aili' early warning machine in 2012 is scrapped after attack; 3 frame Sabo-2000 Aili early warning machine delivery in 2019. And the number of the '10 buckstan active early warning machines' is the same as that of the early warning machine libraries.

Thus, the conclusion is finally: the model hangar is a special hangar for the 2-type early warning machines of the active service of the Pakistan air force, and 8 air force bases are distributed and deployed by 10 early warning machines.

In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (5)

1. The equipment target interpretation method based on the high-resolution remote sensing image is characterized by comprising the following steps of:

step 1: the method comprises the steps of obtaining an original remote sensing image, preprocessing a full-color spectrum image and a multispectral image in the original remote sensing image, and then fusing to obtain a high-resolution remote sensing image;

step 2: screening the remote sensing images according to a set rule, and adjusting the remote sensing images which do not meet the rule to obtain a target analysis high-resolution remote sensing image;

step 3: establishing an equipment target classification system according to the types of the equipment targets;

step 4: extracting characteristic information of the equipment target on the target analysis high-resolution remote sensing image, selecting identification characteristics from the information knowledge, comparing the identification characteristics with the equipment target classification system, and directly interpreting the equipment target to obtain a target primary identification result;

step 5: extracting image information on the target analysis high-resolution remote sensing image, selecting secondary interpretation recognition features from the image information, acquiring surrounding related event information, and indirectly interpreting the image information according to the target primary recognition result, the secondary interpretation recognition features, the equipment target classification system and the surrounding related event information by a logical reasoning method to realize secondary interpretation recognition of the equipment target, so as to obtain a target model;

step 6: and (3) acquiring parameter indexes and target related event information of the equipment targets corresponding to the target model, comparing the parameter indexes and the target related event information with the characteristic information and the image information of the equipment targets, searching contradiction points, returning to the step (5) if the contradiction points exist, and otherwise, outputting the current target model.

2. The method according to claim 1, wherein in the step 1, the multispectral image is subjected to radiation calibration, atmospheric correction and orthographic correction preprocessing, and the full-color spectral image is subjected to radiation calibration and orthographic correction preprocessing.

3. The method for interpreting an equipment target based on a high-resolution remote sensing image according to claim 1, wherein said rule comprises: the spatial resolution of the image of the research area is better than 0.5 meter; the targets of the research area are free of cloud coverage; the side swing angle of the satellite is not more than +/-15 degrees during image acquisition; the adjustment includes image mosaic, image cropping, image band combining, and exposure and color adjustment.

4. The method according to claim 1, wherein the target image shape, size, hue, shade, position and activity of the equipment target are selected from the feature information as the recognition features for recognizing and judging the nature and condition of the target.

5. The method for interpreting equipment targets based on high-resolution remote sensing images according to claim 1, wherein the specific steps of performing the indirect interpretation in the step 5 are:

step 51: extracting the image information on the target analysis high-resolution remote sensing image, and carrying out primary classification according to the target primary identification result to obtain a target large class type;

step 52: extracting the secondary interpretation recognition features from the image information, carrying out secondary classification by combining the secondary interpretation recognition features, the target major class type and the equipment target classification system, and reasoning to obtain a secondary attribute classification result;

step 53: and acquiring the surrounding related event information related to the image information, and acquiring the target model according to the target major type, the secondary attribute classification result, the equipment target classification system and the surrounding related event information.